Load matched raw and DeepInterpolation-denoised fluorescence traces from Allen Brain Observatory Visual Coding 2P sessions (BrainObservatoryCache, Slc17a7 and Cux2-CreERT2 lines, VISp, natural movie one stimulus). Compute participation ratio (PR = (sum lambda_i)^2 / sum lambda_i^2) and effective dimensionality on trial-averaged and single-trial population covariance matrices. Compare PR distributions before and after denoising across N >= 10 sessions, reporting paired effect sizes and 95% bootstrap CIs. Include shuffle controls (identity-shuffled trial labels) to set noise floor. Success criterion: paired t-test on PR detects >10% reduction post-denoising with p < 0.05 FDR-corrected across sessions.
Details
- disease
- none
- target_ref
- Allen Brain Observatory Visual Coding 2P
- primary_endpoint
- Decision brief names ≥3 distinct open questions with at least one supporting citation per question
- identification_strategy
- observational
Raw fields (3)
- assay_spec
Literature triangulation: cross-area latency hierarchy scoring from Siegle 2021 (Neuropixels Visual Coding) vs MICrONS functional recordings — synthesize published hierarchy metrics, identify convergence/divergence, and produce a circuit observability decision brief with ranked open questions; {'kind': 'claim', 'text_template': 'Siegle 2021 hierarchy index ranks visual areas V1 < LM < AL/RL < PM < AM based on inter-area spike-latency and response-reliability metrics from Neuropixels simultaneous recordings across N≥6 areas'}; {'kind': 'claim', 'text_template': 'MICrONS functional recordings [pending verification] either recover or contradict the Siegle 2021 ordering — the direction and magnitude of any reordering constitutes the primary open question for circuit observability'}; Table of hierarchy scores per visual area from Siegle 2021: area label, hierarchy index value, N sessions, N units passing QC (presence ratio > 0.9, ISI violation < 0.5), cross-validation method; Gap analysis table: which hierarchy comparisons are supported by ≥2 independent datasets vs. single-source only; which areas lack task-engaged recordings; which claims require MICrONS reconciliation; circuit_observability_decision_brief — ranked open questions (cross-area latency, DeepInterpolation × geometry, behavior × coding) with evidence strength, tractability, and recommended next experiment per question- hypothesis
Applying DeepInterpolation denoising to Allen Brain Observatory 2P calcium imaging sessions (Visual Coding, natural movie one) will reduce estimated participation ratio and apparent dimensionality of V1 population activity by removing independent noise dimensions, bringing geometry estimates closer to those expected from shared signal subspaces.
- kill_criteria
Abort or revise if: MICrONS 2025 DOI unverified — if full text unavailable, MICrONS claims must be held as TODO_VERIFY and the brief scoped to Siegle 2021 + de Vries 2020 only; Hierarchy index values in Siegle 2021 may be figures-only (not tabulated in text) — extraction may require figure reading rather than text parsing; Rate-limiting on Semantic Scholar from prior tick may recur; fall back to PubMed + direct DOI fetch